基于二次迭代Cr分量的改进Otsu自适应阴影检测算法

城市航空遥感影像中的阴影会造成影像信息丢失和信息干扰,现有阴影检测方法存在误检率高、精度低等问题。提出一种二次迭代Cr分量与改进的Otsu自适应阈值相结合的高分辨率城市航空影像阴影检测算法。将影像两次转换到YCb Cr空间,对其Cr通道直方图进行改进的Otsu自适应阈值选择,确定阴影区域。与二次迭代HSV空间检测方法进行对比实验,所提方法解决了前者的H分量损失和原始Otsu阈值算法选取不准确的问题,在准确率和精度方面具有明显优势。

Improved Otsu adaptive algorithm for shadow detection based on second iterative Cr component

In the color urban aerial images, shadows generated by buildings block the information on the image, resulting in image degradation phenomenon and create difficulties in many application. Thus, it is necessary to effectively detected and treatment to shadows prior to further processing. In this paper, we proposed a method for high-resolution urban aerial image to detect shadows accurately which is based on second iterative Cr component and modified Otsu adaptive threshold. Firstly convert images from the RGB color space to YCbCr color space, then detect shadows with an improved Otsu algorithm in the Cr component histogram.This method solves the problem of inaccurate manual threshold selection. Comparing with HSV color model method, experimental results show that our algorithm improves the accuracy and precision of the shadow detection,and it has a high robustness for small shadows detection.